By Eduardo Pilo de la Fuente,

Sudip K. Mazumder, and

Ignacio Gonzlez Franco

Railway Electrical
Smart Grids
n the last decade, the development of next-
generation electrical smart grids (ESGs) has been one of the priorities in the field of electrical engineering, both for most of the
research centers and for the industry. In short, an ESG consists
of the integration of information technologies into the electrical
system to improve its controllability. In traditional power systems, the control
has been carried out only by the power plants and some elements of the grid
(transformer tap-changers, compensation capacitors, and reactances), whereas,
in the next-generation smart grids, most of the elements can respond to control
orders from the system operator, which allows, for instance, for the integration
of the distributed generation and the demand into the control schemes of the
power system. Because of this improved controllability, ESG technologies promise a significant improvement in the capacity utilization, the reliability of the system, and the energy efficiency of the grid.
Although rail power systems (RPSs) are a special case of electrical power
system, they are operated in a very different way. While the goal of the power

An introduction to
next-generation
railway power systems
and their operation.

Digital Object Identifier 10.1109/MELE.2014.2338411

Date of publication: 29 September 2014
Substation Image: courtesy of
Eduardo Pilo de la Fuente,
Binary numbers Microsoft

2325-5987/142014IEEE

IEEE Elec trific ation Magazine / s ep t emb er 201 4

49

system is to provide electrical power with the agreed characteristics, the final objective of the railway system is to transport passengers and goods according to a schedule. For that
reason, the operation of a railway ESG (RESG) has to be different from conventional ESGs.
Most of the RPSs share some characteristics that make
the development of specific ESG technologies adapted to
railways particularly important.
1) Most electrical loads are trains, which are spatiotemporally varying loads. The consumption of the trains is
related to the way each train is driven, which allows it
to even become a generator for a limited amount of
time when braking. Just by stopping acceleration and
starting braking, the load can vary from 10 to 8 MW in
a few seconds. The RPS has to be able to deal with these
changes, which occur very often along a given journey.
2) Electrified railways are normally considered one of the
most energy-efficient modes of transport, especially
over economically viable operating distances. Its potential for energy savings is largely due to regenerative
braking, whose efficiency depends largely on when and
where it is carried out.
3) Railway lines normally cross wide areas and, therefore, are often interconnected to several electrical
grids, which are normally heterogeneous, as strong
grids coexist in the field with weaker grids. A smart
control that takes into account the specificities of
each network is crucial to improve the overall reliability and capacity utilization.
4) An RESG relies heavily on good bidirectional
communications between trains and the infrastructure, which is sometimes difficult to achieve, for
instance, in tunnels or in remote areas.

ESO 1

This article describes railway power systems and their

operation. In addition, the main control actions that can be
performed by an RESG are introduced, explaining how they
can improve the performance of traditional RPSs, e.g., reducing
costs, increasing energy efficiency, and enhancing reliability.

Railway Power System Grids

System Description
As shown in Figure 1, RPSs normally take the electricity
from other power systems, which, in turn, have their own
generation plants and electrical grids (transmission grid for
bulk power transfer, and distribution grid for retail power
supply) and whose characteristics may vary significantly
(strong grids coexist in the field with much weaker grids).
In liberalized electricity sectors, the transmission, distribution, and generation activities are typically carried
out by different companies. The electrical system operators (ESOs) are the companies in charge of balancing generation and demand and operating the transmission grid
in such a way that the reliability of the system is guaranteed. The distribution system operators are companies
that operate the distribution grids in such a way to ensure
that electricity is supplied to every customer with the
required quality. Finally, the generation companies are
responsible for producing the energy that has been programmed in each power plant. The energy produced can be
sold by means of contractual agreements or in organized
electricity markets (generally spot markets, including dayahead and intraday sessions), operated by an electricity market operator. However, final corrections to the program are
introduced by the ESO to solve technical restrictions and to
respond to the unexpected variations that occur in real time.

ESO 2

EMO 1

Power Plants

Power Plants

T&D Grid1
Power System
Side
Traction
Substation
Railway
Power System

Legend
EMO: Electricity Market Operator
RDG: Railway-Side Distributed Generation

Traction
Substation

I E E E E l e c t ri f i c a t i on M a gaz ine / september 2014

T&D Grid 2

Traction
Substation
ESS

Figure 1. The interconnections of railway power systems to other power systems.

50

EMO 2

RDG

Traction
Substation

Traction
Substation

In some countries, some transmission and distribution

(T&D) grids and power plants, owned by railway companies,
are used specifically for traction purposes. Because of their
usage, these grids have particular characteristics (e.g., twowire, single-phase lines using a low frequency of 16.6 Hz).
Regarding the railways side, when it is not vertically
integrated, the infrastructure manager [referred to as the
railway system operator (RSO) as an analogy with ESOs in
power systems] plays a dual role: 1) regulating the traffic
to ensure safety and an adequate flow of trains and
2) controlling the railway power system. The train operators are independent companies whose activity is to
transport loads or passengers with trains. As users of the
railway infrastructure, they must pay for the services they
use, including power supply, to the RSO.
As shown in Figure 1, RPSs are connected to transmission or distribution grids by means of traction substations
(TSSs). It should be noted that not all TSSs are fed directly
from the T&D grid: sometimes railway-side electrical lines
connect several TSS. In the case of dc-fed railways, substations include transformers and rectifiers. In the case of acfed railways, substations include mainly transformers
and, when the frequency of the T&D grid and railways is
different, frequency converters (static or rotary).
Although they are not widely used, energy storage systems (ESSs) allow the temporal excesses of power to be
stored and used later in a deferred way.
Finally, the railway power plants (referred to as railwayside distributed generation (RDG) in Figure 1) are generators
(typically distributed sources of energy) controlled directly
by the RSO, which allows for railway-oriented operation,
and connected directly to the railway grid.

The Operation of Electrified Railways

Control of the System
The operation of an electrified railway includes two different facets that have to be controlled in a compatible way:
1) the traffic flow operation (which refers to the way the
trains move) and 2) the electrification operation (which
refers to the way the power is supplied). Therefore, the
control centers of the railway are charged with supervising and operating both the traffic and the electrification.
To ensure the safe and efficient operation of the railway,
a signaling system is typically in place to manage the traffic flow. The traffic control is typically structured in layers.
First, a protection layer is responsible for the safety of the
train movements and is in charge of giving the orders to
ensure that no train leaves its safe operation conditions
(for instance, by getting too close to another train or by
exceeding its maximum speed in a specific section).
Depending on the specific technology used in the signaling
system, the degree of automation of the control may be
very different: from manual control (based on visual signals and relying on a person taking the right actions) to
fully automated control (based on communications and

relying on a control unit to ensure that the system is always

in a safe state). Additional layers, always subordinated to the
protection layer, are commonly used to improve the quality of the traffic flow according to different criteria (such as
punctuality and regularity). The control related to energy
consumption optimization would correspond to these
operational layers.
The electrification has a similar architecture. A first
layer, in charge of protecting the electrical equipment and
infrastructure, continuously checks if all of the electrical
quantities (voltages, currents, etc.) are within the allowable
range and, otherwise, isolates the failure to avoid further
damages. Additional layers are responsible for optimizing
the operation of the railway grid by reconfiguring its topology, operating the tap changers of the transformers, etc.
Unfortunately, these control actions are quite limited and
are often too slow for launching them frequently. Thus, the
grid is normally designed to supply power in the worst-case
scenario with very few control actions, which leads to quite
oversized infrastructures.
Although the traffic and electrification are two facets
physically coupled in railways (the electrical loads depend
on the way each train is driven and the way a train is driven
depends on the voltages and, therefore, on the electrical
loads), even the upper layers of these two control systems
are usually completely uncoupled.

The Operation of Train Services

An important concept for understanding RPS operations is the
interrelation between the train movement and its power consumption, which can be used to accelerate the train, to compensate the losses due to running resistance forcesred curves
in Figure 2and/or to feed the onboard equipment (air conditioning, pumps, compressors, lighting, etc.). Similarly, when a
train equipped with electrical braking systems brakes, the
kinetic energy is converted into electrical power and used to
feed onboard equipment, to feed other electrical loads by injecting this power back into the catenary (regenerative brake), or, if
none of the previous options is possible, to heat up the onboard
resistors installed for that purpose (rheostatic brake).
The power usage related to train movement depends
essentially on how the train is driven. The driver, which can be
a person or an automatic driving system, decides which force
is required to move the train as wanted within the operating
limits of the train (see Figure 2)this depends on the voltage
and the speed at which it is operating. Four types of driving
actions are normally performed: 1) accelerating, where the
traction equipment exerts a force to increase the speed, 2)
braking, by exerting a force to reduce the speed, 3) cruising, by
exerting only the force required to compensate the running
resistance (which maintains the speed), and 4) coasting, when
the train does not exert any force at all.
For a given journey duration, a train can be driven in
many different ways: accelerating, braking, and coasting differently (in different locations and with different intensities).
One driving strategy commonly used for analysis is the

IEEE Elec trific ation Magazine / s ep t emb er 201 4

51

300
kN
250

SERIE 103 Tare Weight: 425 t; Power: 8,800 kW;

Maximum Speed: 350 km/h
Traction 100%
(8,800 kW)
Traction 75%
(6,600 kW)

200
150

Traction 50%
(4,400 kW)

100
50
0

Ramping Drag 25 mm
Horizontal Drag
50

100

150

200
km/h

250

300

350

50
Rheostatic Brake
100

150

200

250

Regenerative
Brake with
Maximum
Power
(8,800 kW)

300
Figure 2. The maximum traction and braking forces for series 103
trains from Renfe, Spain. (Figure courtesy of Luis E. Mesa.)

inimum time driving (MTD), which

m
corresponds to the fastest way of driving while satisfying the limits of the rolling stock and the infrastructure. The
commercial driving is normally
designed by adding some time margins
to the MTD to allow the trains to
respond to the small perturbations that
occur in real operation (typically delays).
It is important to highlight that
each driving style can lead to a very
different spatiotemporal distribution
of the power consumptions (see Figure
3) and, consequently, to significantly
different requirements for the RPS.
This flexibility is the key for conceiving smart strategies for driving the
trains for many different purposes,
such as saving energy and augmenting the traffic capacity.

52

I E E E E l e c t ri f i c a t i on M a gaz ine / september 2014

To manage the traffic flow, each train has to fulfill a

schedule, which means that it must reach the next station
[or the next regulation point (RP)] in the specified time
(see Figure 4), within a tolerance (represented in green).
The time in these intermediate RPs is checked to allow for
a better adjustment of the train driving: if the train arrives
too early, it can slow down or, conversely, drive faster if it
is delayed.
In the last decade, many researchers have worked intensively to design ecodriving strategies, i.e., driving strategies
that minimize the energy consumption of the train. Important energy savings have been achieved by smartly adjusting
the coasting sections, which avoids consuming energy that
will be given back to the catenary (or wasted in rheostats)
later. However, coasting typically augments the trip duration,
and, therefore, a tradeoff between energy savings and trip
duration has to be found by using o
ptimization techniques.
The design of schedules optimized to enhance operation has also been a very active research topic. For
instance, in dc railways, such as commuter trains and
metros, the synchronization of departures, arrivals, and
driving strategies has been used to improve the receptivity
of the contact line (catenary or active rail) when regenerative braking is used, leading to significant energy savings.
There are also experiences in which similar techniques
have been used to improve the traffic capacity of a line.
Both techniques, ecodriving and the smart scheduling,
have been very successful in improving the performance
of electrical railways, especially in terms of energy efficiency, when the operation is planned. As they require a
huge computational effort, their application to real-time
control is very unusual today.

The Next-Generation Railway ESGs

Although improving energy efficiency has often been
claimed as the main change vector toward the future
railway ESGs, it is important to highlight the enhancement of the controllability of the electrified railways that
can be achieved with RESG technologies. By managing the traffic and electrification in an integrated way, the
RESG can efficiently solve different
operation problems (capacity limitations, changes in the planned operation, etc.), including many issues that
cannot be addressed within the traditional control schemes. To allow this,
the RESG has to integrate the missions of the railway (to move trains to
transport goods and persons) and the
electrification (to supply the electricity required by the trains), as represented in Figure 5.
To explain why the RESG can
improve the operation of electrified

By managing the
traffic and
electrification in an
integrated way, the
RESG can efficiently
solve different
operation problems
that cannot be
addressed within the
traditional control
schemes.

railways, it is important to underline

Maximum Speed
that RPSs are generally not infinite
Maximum Speed
grids. RPSs are normally designed to be
able to supply electrical power in the
Speed
Speed
worst-case conditions defined in the
requirements, both for normal operation (with all of the elements of the
Time
Time
system working properly) and for
degraded operation (for instance,
assuming the loss of one or two subPower
Power
stations). In the design process, a specific operation is assumed, including
Figure 3. Two examples of different driving strategies for the same trip and duration (single
rolling stock characteristics, train fre- train), leading to different power profiles.
quencies, and driving strategy (typically MTD). Once the electrification is in service, the operation
to address, especially if a representation of the electrificaconditions change as time goes on: the transport demand
tion is included in the optimization model. It should be
tends to grow (following the economy growth) and so do the
noted that the main objective of the STD can be very
electrical requirements. As long as the operation is less
diverse: minimizing the energy consumption, adapting
demanding than planned, the electrification can provide the
the train consumption to the capacity of the infrastructure
power the trains request. But once the limits of the electrifiin a specific area, and reducing the cost of the electricity.
cation start being reached (e.g., some line sections are temIn Figure 6, in addition to minimizing the energy conporally overloaded, the voltage drops become too large in
sumption (case A), which has been taken as the base case,
specific points, etc.), the electrification starts creating bottletwo other types of driving changes are introduced. Case B
necks to the operation at specific peak moments in specific
corresponds to a limitation of the power peak supplied by
locations. When this occurs, it is, of course, possible to
the electrical grid 2, e.g., due to temporary capacity limitaupgrade the electrification by adding some reinforcements
tions. Naturally, depending on their type (current or volt(additional conductors to the catenary, new substations, etc.).
age capacity limitations), the power consumption should
A different approach would be possible by using RESG techbe modulated differently for better results. If these limitanologies: adapting the operation so that the rated limits of
tions were at a TSS level (instead of at an electrical grid
the electrification are not exceeded. This is the goal of several
level), the adjustments would be similar, but covering a
ongoing research projects, which are exploring these control
different area. Finally, case C corresponds to a transfer of
mechanisms and developing different RESG technologies. An
part of the energy consumption from electrical grid 2 to
example is the project MERLIN, a European initiative that is
electrical grids 1 and 3, which could be advantageous, for
expected to deliver the final results by the end of 2015 (see
instance, if the prices of the energy were higher in the
The MERLIN Project).
electrical grid 2 [price-oriented driving (POD)].
This section describes some of the features that will be
It should be noted that, in general, modifying the power
possible in future RESGs, grouped into three categories:
profile normally implies modifying the speed profiles and,
1) smart train operation, 2) smart operation of the RPS, and
therefore, the arrival times to the RP. Consequently, in addi3) smart interaction with other power systems.
tion to a spatial shifting of the power consumption, a temporal shifting is also performed. When performing POD, this
Smart Train Operation
can be useful as electricity prices often vary, not only with
As discussed in the Railway Power System Grids section,
the location of the supply but also with the time.
train driving provides a flexible tool to the RSO to adjust the
In addition to driving the trains, another important aspect
power consumption profiles to the needs of the system. As
of the smart train operation is the management of all of the
the traction energy consumption represents an important
part of the operation cost, railway companies have devoted
Position
much effort to optimize the driving strategies to minimize
RP 4
the energy consumption. This is normally an offline
(Destination)
process where the results are a reduced set of driving stratRP 3
egies, each for a different journey duration.
Train 1
Train 2
Smart train driving (STD), i.e., controlling online the
RP 2
way the trains are driven, is the most direct mechanism
for performing an active management of the demand
RP 1
Time
(AMD), a key feature in most ESGs, in RPSs. Because of the
(Origin)
computational load it involves, optimizing the train drivFigure 4. The traffic mesh for two consecutive identical trains.
ing in a short time is a major challenge the RESG will have

IEEE Elec trific ation Magazine / s ep t emb er 201 4

53

Traction
Substation
Railway
Power System
(Controlled by
RSO)

Traction
Substation

Traction
Substation
ESS

Railway
System
Operator

Traction
Substation

RSO

RDG

Controls the
Railways Power
Flows ( ) and
the Traffic Flow
(
)

Figure 5. The control of future railway ESGs.

auxiliary loads onboard (e.g., the cooling system of the

traction equipment, air conditioning, lighting, and entertainment equipment). Some of these loads can be managed in
smart way, modulating the power consumption according to
the needs of the system (and the train itself).

Smart Operation of the RPS

RPSs normally have a limited set of controllable devices,
including the switching devices (breakers, disconnectors,
etc.), on-load tap changers of the transformers, and converters used to connect the RPS to the T&D grid (in 16.6-Hz
systems and similar). Because of this limited controllability, the conception of AMD strategies has been, so far, the
most common approach in RESG research.
However, this tendency is very likely to change in the
future thanks to the advances in power electronics that

make it possible to reach higher voltages, transferring more

power more efficiently and more affordably. With these
technologies, a smart control of the power flows within the
RPS would be possible.
Figure 7 compares two different concepts for controlling the power flows. In solution A [Figure 7(a)], the power
flow is controlled at every coupling point to the supplying
grids (in the TSS) by the controlling devices labeled PFR.
Alternatively, in solution B [Figure 7(b)], the controlling
devices (PFE in the figure) set the power flows between
adjacent sections, which indirectly control the power flow
supplied by each TSS. (The ESS and RDG are controlled in
the same way as in solution A.)
Regardless of which solution is adopted, controlling the
power flows within the RPS can significantly help to
achieve the purposes of the RESGs, such as minimizing the

The MERLIN Project

54

he MERLIN project (http://www.

Network Rail, and Trafikverket), supported

leading to a cost-effective, intelligent

merlin-rail.eu) is an important initiative

by consulting companies (DAppolonia),

management of energy and resources through:

in the European Union (EU) context that

universities (Newcastle University and

comes up as a response to the fifth call

RWTH-Aachen University), and research

issued by the European Commission as

centers and professional associations [the

part of Seven Framework Programme.

Association of European Railway Industries

The partnership gathered to achieve the

(UNIFE), the Union Internationale des

railway operations and procedures on

MERLIN objectives is composed of 20

Chemins de Fer (UIC), and the Spanish

energy demand

partners from eight EU member states

Railways Foundation (FFE)].

improved design of railway distribution

networks and electrical systems and
their interfaces

better understanding of the influence of

identification of energy usage optimizing

technologies

(Czech Republic, Ger many, France,

As mentioned on its official site, the aim

Belgium, Italy, Spain, United Kingdom,

of the MERLIN project is to investigate and

improved traction energy supply

and Sweden), comprising different

demonstrate the viability of an integrated

understanding of the cross-dependencies

European railway systems integrators and

management system to achieve a more

equipment suppliers (ALSTOM, AnsaldoSTS,

sustainable and optimized energy usage in

AnsaldoBreda, MerMec, SIEMENS, CAF, and

European electric mainline railway systems

Oltis Group), along with railway operators

and to provide an integrated and optimized

(RENFE), infrastructure managers (ADIF, RFF,

approach to support operational decisions,

I E E E E l e c t ri f i c a t i on M a gaz ine / september 2014

between technological solutions

improving cost-effectiveness of the

overall railway system

contribution to European standardization

(TecRec).

Electrical
Grid 2

Electrical
Grid 3

Electrical
Grid 1

(a)

Electrical
Grid 1

Electrical
Grid 3

Original
Modified

Position

Original

Position

(b)

Electrical
Grid 2

Electrical
Grid 3

TSS TSS TSS TSS TSS TSS

TSS TSS TSS TSS TSS TSS

Power

Power

TSS TSS TSS TSS TSS TSS

Electrical
Grid 2

Power

Electrical
Grid 1

Modified

Position

(c)

Figure 6. The different driving changes executed to manage the power demand in an RPS: (a) base case, (b) power peak reduction, and (c) energy transfer.

Legend
PFR: Power Flow Regulator
PFE: Power Flow Equalizer
: Controlling Device
Electrical Power
Systems
Railway Power
System

Traction
Substation

Traction
Substation

Traction
Substation

Traction
Substation

PFR

PFR

PFR

PFR

(a)
Electrical Power
Systems
Railway Power
System

Traction
Substation

Traction
Substation

PFE

Traction
Substation

PFE

Traction
Substation

PFE

(b)
Figure 7. The different strategies to control the power flows in an RPS: (a) case A: control at the power supply points and (b) case B: power equalization.

losses, reducing the cost of the electricity, and smoothening

the voltage profiles. But, in addition, as the power flows can
be readjusted to come from different TSSs, the reliability of
the system is significantly improved. In some cases, this
may allow for a reduction of the overrating of the infrastructure elements. (Transformers, converters, and lines are
designed to work in conditions that could be avoided or
mitigated by using RESG.)

Smart Interaction with Other Power Systems

Probably the most important aspect of the RESGs is the
improvement of controllability that can be achieved, which
makes it possible to adapt the operation in real time to the
oncoming events originated inside or outside the domain
of the RPS. Because of their relative size, RPSs have an
impact on the T&D grids to which they are connected,
which has to be carefully analyzed to avoid nuisances to
other customers. But, for the same reason, they can also
efficiently help the ESOs and the T&D grid operator perform
an appropriate operation.

Here are just two examples of the richer interaction

between heterogeneous smart grids (RESGs and ESGs) that
will be possible with the adoption of RESG technologies:
xx
When an incident occurs in the T&D grid and its
capacity has been reduced temporarily, the T&D grid
operators can prioritize other customers and ask the
railway to reduce its consumption from a specific set
of substations: the RESG allows it.
xx
With RESG technologies, in the future, railways could
also provide ancillary services (e.g., secondary band
regulation) to help balance the generation and the
demand in an electrical system.

Biographies
Eduardo Pilo de la Fuente (eduardo.pilo@eprail.com) is
with EPRail Research and Consulting, Spain.
Sudip K. Mazumder (mazumder@uic.edu) is with the
University of Illinois at Chicago.
Ignacio Gonzlez Franco (igonzalez@ffe.es) is with the
Spanish Railways Foundation (FFE).

IEEE Elec trific ation Magazine / s ep t emb er 201 4

55